Paris's War on Duplicate Images in Public Archives: What Officials, Experts and Key Figures Are Saying

Paris's network of public digital archives is sitting on a problem that archivists have quietly flagged for years but that officials are now being pushed to address openly: tens of thousands of duplicate images stored across municipal and national repositories, inflating storage costs, slowing retrieval systems and undermining the integrity of records that the city's urban-regeneration programs depend on.

The issue has gained urgency in 2026 because of the Paris 2024 Olympics legacy activation programme, which requires city agencies to draw on photographic and cartographic records stretching back to the 2015 candidacy period. Planners at Aménagement de Paris, the municipal urban-development arm, have begun auditing image databases tied to the Grand Paris Express metro expansion—a project whose documentation alone runs to hundreds of thousands of files distributed across at least four separate agency servers along the future Line 15 corridor.

Why the Problem Is Getting Worse

Digital duplication in public archives is not new, but the scale has grown sharply since 2020, when Paris accelerated its transition to cloud-based document management following the Covid-19 lockdowns. The Bibliothèque nationale de France, on the Quai François-Mauriac in the 13th arrondissement, manages one of Europe's largest public image collections and has internally acknowledged the challenge of deduplication at institutional scale, according to published annual reports from the BnF covering the 2022–2024 period. Storage infrastructure costs for large public archives in France have risen alongside those pressures.

At the city level, the problem is more diffuse. The Direction de l'Urbanisme at the Hôtel de Ville coordinates planning imagery across a patchwork of systems—some managed by the Atelier Parisien d'Urbanisme, known as APUR, others held by individual arrondissement administrations. Specialists in digital asset management point out that when agencies digitise the same physical document independently, duplication is almost automatic without a shared deduplication protocol in place from the start.

Experts in the field argue that the technical fix is well understood: perceptual hashing algorithms can identify near-identical images even when file names or metadata differ. The harder question is governance—specifically, which agency takes responsibility for running deduplication across systems that belong to different administrative layers of the French state.

What Key Figures Are Calling For

The debate has drawn in figures from several corners of Parisian public life. Archivists at the Archives de Paris, housed on the Boulevard Sérurier in the 19th arrondissement, have publicly advocated in professional journals for a centrally mandated deduplication standard tied to the broader interoperability framework that the French government's Direction interministérielle du numérique, DINUM, has been developing since 2021. That framework, known as the Référentiel Général d'Interopérabilité, sets technical standards for public digital systems but does not currently include specific provisions for image deduplication.

Researchers at Sciences Po's urban-studies department have flagged the downstream consequences for policy. When planners working on Seine river-bank regeneration between the Pont de Bercy and the Pont de Tolbiac pull historical image sets to track waterfront change over time, duplicate files distort automated analysis—a batch of aerial photographs tagged to 2019 may contain the same image filed three or four times under different project codes, skewing any algorithmic assessment of change rates.

The pressure is also financial. Cloud storage is not free, and municipal budget constraints under the current Macron second-term governance have sharpened scrutiny of avoidable IT expenditure across all public bodies. Industry benchmarks cited in a 2024 report by the Observatoire du Numérique Public suggest that unmanaged duplication can account for 20 to 30 percent of an institution's total image storage volume.

What happens next depends largely on whether DINUM moves to extend its interoperability framework to cover image asset management before the end of 2026—a timeline that archivists and urban planners both say would align with the ongoing Grand Paris Express documentation cycle. For institutions that cannot wait, APUR has begun piloting an in-house deduplication workflow on its Seine regeneration image library, a model that smaller arrondissement-level agencies are watching closely before committing to their own system upgrades.